肌肉组织背景下的微血管分割方法研究

微血管的分割提取是微循环网络特征参数分析与功能障碍评估的关键性图像处理技术基础之一。目前尚缺乏有效的专门针对肌肉组织来源图像的微血管分割方法,使用现有的血管分割算法难以取得理想的分割效果。本研究依据该类图像的特点,针对性地提出了一种基于多技术组合联用的分割方案—TOSC分割法。首先以阈值法进行初次分割,然后通过形态学开运算截断背景与微血管间连接区,再以种子填充法对微血管区域进行二次识别与分割,最后通过形态学闭运算填充分割区域中的孔洞和缺刻,实现了微血管的分割提取以及在此基础上的功能微血管面积密度测量。分割测试结果表明TOSC分割法能够准确地将微血管区域从复杂的肌纤维组织背景中提取出来,可满足基础研究与临床处理的要求。

Microvascular Segmentation from Image Background of Muscular Tissue

The segmentation of microvessel is one of critical image processing techniques to evaluate property and dysfunction of microvasculature.Up to the present,has no specific segmentation method been presented to deal with microvessels in muscular tissue images.Presented segmentation methods for vessels are not competent for this type of images.In this study,we proposed a combinatorial strategy — "TOSC algorithm".first,segment the image preliminarily with single threshold was segmented;second,the connections between background and microvesselswas splited by morphological opening operation;third,secondary recognition and segmentation were executed by seed fill algorithm;then,the lacunas was filled by morphological closing operation.Vascular area was extracted and functional microvascular density was calculated.Testing proves that TOSC algorithm can segment and extract microvascular area accurately from cremaster muscle images.